The standard Arduino printing of floats only prints if its value fits in an uint32_t (4.294.967.295). Above that it will print ovf for overflow.
The printHelpers library supports the printing in scientific format e.g. 3.897E12 to solve this artificial limit.